The present invention relates to a solid state imaging device including a photosensitive pixel for monitoring, and applicable to an automatic focusing device for a camera and the like.
A charge coupled device (CCD) image sensor has been well known as a solid state imaging device. There has been also well known a CCD image sensor including a photosensitive pixel for monitoring, which is applicable to the sensor used in an automatic focusing device of the camera and like, in order to control a storage time of a signal charge in each photosensitive pixel (for a picture) and a gain of an output circuit which converts the signal charge into an output signal.
FIG. 1 is a schematic view showing an example of a conventional CCD image sensor. In FIG. 1 a plurality of photosensitive pixels for a picture are arranged in a matrix shape on a semiconductor substrate (not shown) so as to form a plurality of pixel trains P.sub.1 through P.sub.n. A photosensitive pixel 1 for a monitor generates signal charges in proportion to a mean value of an incident light amount of the photosensitive pixels for a picture.
The signal charges generated in the photosensitive pixel 1 for a monitor are outputted as a signal V.sub.AGC through an output buffer 2 so as to supply to a control circuit (not shown) externally provided. The control circuit opens a shift gate S at each time when the supplied signal V.sub.AGC reaches a predetermined threshold value, so as to transfer the signal charges to a charge coupled device (CCD) register 4. Each of the signal charges transferred into the CCD register 4 are converted into an output signal in an output circuit 5 so as to be externally outputted. Here, the output circuit 5 comprises an output buffer 8, a voltage amplifying circuit 7, and a clamp circuit 8. The voltage amplifying circuit 7 changes over a gain according to an instruction from the control circuit when the signal V.sub.AGC is less than the threshold value, for example, because an incident light amount is little. Furthermore, the clamp circuit 8 causes the output signal of the output circuit 5 to be within an input range of an analog/digital (A/D) converter provided in the latter stage, in which the lower and upper most values are set by the control circuit.
In this manner, in the CCD image sensor as shown in FIG. 1, the monitor photosensitive pixel 1 is provided as means for measuring the mean value of the signal charges of each picture photosensitive pixel constituting the pixel trains P.sub.1 -P.sub.n.
However, since the conventional CCD image sensor includes the output circuit 5 having the output signal of the output buffer 2 and the output signal V.sub.AGC which have different voltage gain and off-set voltage, the monitor photosensitive pixel 1 has a very bad accuracy as the measuring means for the mean value as described above.
Therefore, the CCD image sensor as shown in FIG. 1 also has the bad accuracy at the time when the threshold value is determined at opening the shift gate 3. Accordingly, such the CCD image sensor can not set the optimum threshold value corresponding to the incident light amount.
Furthermore, since the accuracy of the output signal V.sub.AGC is wrong, it is impossible to set the optimum values of the gain of the voltage amplifying circuit 7 and the upper and lower most values of the clamp circuit 8 by using the output signal V.sub.AGC. Therefore, it has been on general to set the gain of the voltage amplifying circuit 7 and the upper and lower most values of the clamp circuit 8 to be the optimum value by using as a reference value the signal charges which are generated in a shaded pixel provided at the head of the pixel trains P.sub.1 -P.sub.n and which is a sinai pixel when the incident light amount is zero.
However, in the method of using the signal charges of such the shaded pixel, since the lowest level is used as the reference level without the means level, it is impossible to sufficiently and accurately set the gain of the voltage amplifying circuit 7 and the upper and lower most values of the clamp circuit 8. Such a fault becomes obvious according to the case where the contrast is small in the incident light amount of the pixel trains P.sub.1 -P.sub.n.